Structural requirements in calmodulin for n-methyltransferase recognition and NAD kinase activation

Author

Jennifer Ann Cobb

Date of Award

5-2000

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Biochemistry and Cellular and Molecular Biology

Major Professor

Daniel M. Roberts

Committee Members

Cynthia B. Peterson, Wesley D. Wicks, Frank D. Larimer

Abstract

Calmodulin is trimethylated at lysine 115 by a lysineN-methyltransferase that utilizes S-adenosyl methionine asa co-substrate. Lysine 115 is located within a six aminoacid loop (LGEKLT) between EF hand III and EE hand IV in the carboxyl terminal lobe. Previous studies have shown that the methyltransferase is highly specific for calmodulin, and that minimal requirements for methylation are a folded carboxyl terminal lobe (Han, C.H., Richardson,J., Oh, S.H., and Roberts, D.M. (1993) Biochemistry 32,13974-13980). In the present work a mutagenesis approach was used to investigate the structural features of the carboxyl terminal lobe that are required for calmodulin methylation.Three approaches were taken. First, based on the symmetry between EF hands I and III, and EF hands II andIV, a series of domain (entire EF hand) and subdomain(individual helix or Ca²⁺-binding loop) exchange mutants were generated in which symmetrical elements in the amino terminal lobe were substituted into the carboxyl terminal lobe. Secondly, a converse mutation was made in which the methylation loop sequence was introduced between EF hands I And II in the amino terminal lobe. Finally, a series of site-directed mutations in the methylation loop and adjacent α-helices (helix 6 of EF hand III and helix 7 of EF hand IV) were generated and analyzed.Substitutions at three conserved positions in the methylation loop (G113S, E114A, L116T) abolished the ability of calmodulin to be methylated. A fourth mutant,L112T showed normal methylation in the presence of calcium,but a 4.5 fold reduction in catalytic efficiency in the absence of calcium. These results suggest that conservation of the methylation loop sequence is essential for methylation. However, results with domain mutants CaM^EKL,CaM¹²¹⁴ and CAM¹²³²suggest that the methylation loop sequence alone is not adequate for methylation. CaM¹²³² has the methylation loop sequence placed at a symmetrical position between EF hands I and II in the amino terminal lobe. has EF hand III replaced with EF hand I and CaM¹²³²has EF hand IV replaced with EF hand II. None of the domain exchange mutants were methyltransferase substrates suggesting that structural features unique to both EF handsIII and IV (and not found in EF hands I and II) arerequired.To identify these unique regions in EF hands III andIV, a series of subdomain mutants were constructed. The Most substantial effects were observed with the substitution of helix 2 in EF hand I for helix 6 in EF handIII (CaM^H6) and helix 3 in EF hand II for helix 7 in EF hand IV (CaM^H7). CaM^H6 was a poor methylation,,substrate in both the presence and absence of Ca²⁺, and showed a 20-fold (apo)and 13-fold (Ca²⁺-bound) reduction in catalytic efficiency compared to VU-1. CaM^H7 was not a substrate in the absence of calcium, but showed normal methylation kinetics in the presence of calcium. Thus, the substitution of both helix 6 or 7 affected methyltransferase recognition but in a different fashion, since calcium binding restores the ability of the CaM^H7 to be recognized by the methyltransferase. Helix 7 shows a high number of surface exposed acidic side chains (6 out of 11 residues). To examine the role of these negative charges, a number of calmodulin mutants with substitutions for the various glutamate/aspartate residues were analyzed. Substitutions at residues 118,120, 122, 126,and 127 were found to affect the rates of methylation [with rates ranging from 5% (VU-12) to 60%(D122A) of wildtype]. Similar to CaM^H7, activity with most of these mutants was enhanced in the presence of calcium.In contrast to helix 7, single amino acid substitutions in helix 6 showed no effect on the methylation rate. Thus, the loss of methylation in CaM^H6 is not due to the substitution of any specific amino acid.Rather, it appears that the replacement of the entire sequence of helix 6 with helix 2 resulted in an unexpected change in the conformation that disrupts the binding site for the methyltransferase in both the apo- and calcium-bound states.Overall, the highly conserved methylation loop and α-helices (6 & 7) flanking the loop appear to be important for methyltransferase recognition. However, the role these regions play in methyltransferase recognition appears to be very different. The most critical residues appear to be within the methylation loop (G113, E114, L116). These Residues are very close to site of methylation on lysine-115, and might provide important points of contact with the active site of the enzyme or provide flexibility (e.g.,G113) or stability (L116) for the conformation of the loop structure. Secondly, the conservation of negative charges on helix 7 appears to be important particularly in apo-calmodulin. After the binding of calcium, a hydrophobic surface is exposed and the substitutions on helix 7 appear less important for interaction. This supports previous findings that suggest that ionic forces are important for interaction of apo-CaM with the methyltransferase whereas additional hydrophobic interaction occur with Ca²⁺-CaM (Han,C.H., Richardson, J., Oh, S.H., and Roberts, D.M. (1993)Biochemistry 32, 13974-13980).Lastly, it is clear that helix 6 is important for methyltransferase recognition but it is less certain whether specific interactions with the enzyme take place.It is likely that packing interactions of this helix with others in the carboxyl terminal lobe may be important for stabilizing the conformation of the residues that are recognized and bound by the methyltransferase. Additional Structural work of the calmodulin mutants in question, as well as the methyltransferase, will be necessary to clarify further the specific role of these structural elements.During the course of this study, the activator properties of each calmodulin mutant were also tested with two representative calmodulin-dependent enzymes, NAD kinase(NADK) and cyclic nucleotide phosphodiesterase (PDE).While most mutants activated these target proteins in normal fashion, mutant calmodulins with substitutions in the loop-turn region of the amino terminal lobe and helix 6 of the carboxyl-terminal lobe showed defects in NAD kinase activation. CaM^EKL activated PDE normally, but did not activate NAD kinase and was actually a potent antagonist of the enzyme with a binding affinity similar to wild type calmodulin. X-ray scattering studies support this and showed that the conformation of CaM^EKL in complex with target peptide is similar to wild type calmodulin.While CaM^EKL was a complete NAD kinase antagonist,single substitutions within this region were "partial agonists", activating the enzyme to a fraction of the level observed with wild type. Similarly, it was found that both CaM¹²¹⁴ and CaM^H6were partial agonists, and that the defect of these two calmodulins was the result of a substitution at threonine 110. Overall, these results show that the binding and activation of NAD kinase can be distinguished,and that the residues required for the activation of PDEare distinct from those required for NAD kinase. Further,these regions on calmodulin are within or adjacent to a"latch domain" which is formed between the helix 2 of the amino terminal lobe and helix 6 of carboxyl terminal lobe upon enzyme binding (Meador, W., Means, A.R., and Quiocho,F.A. (1992) Science 251, 1251-1255). This domain has been proposed to be essential for achieving an activated state.

Recommended Citation

Cobb, Jennifer Ann, "Structural requirements in calmodulin for n-methyltransferase recognition and NAD kinase activation. " PhD diss., University of Tennessee, 2000.
https://trace.tennessee.edu/utk_graddiss/8243